It's not that much, but it's also not nothing. The input file of the challenge was very narrow, just a relatively short text column and a numeric value. It still is >13 GB uncompressed; an actual table with a few more columns with 1B rows will easily be beyond 100 GB. And chances are it's not the only table in your database, so it won't fit into RAM of most common database machines. I.e. still a non-trivial dataset in most contexts.